Business & Technology Nexus

We had some friends over for dinner Saturday, and it was the first time I was embarrassed to have my blog entries quoted back to me. It’s definitely a weird experience to throw your opinions out here on these pages & then have your network of friends and acquaintances so “dialed in” to your latest thoughts.

Besides enjoying some out-of-this-world blueberry pie, I was also given a push to go back and dig deeper into the Oracle v. SAP license revenue battle. The idea was to gather Siebel and Peoplesoft revenue pre-acquisition & perhaps over a 2-3 year time horizon, and then essentially chart Orace+Peoplesoft+Siebel v. SAP. The theory is that this chart will give a better read on how the battle is shaping up.

So I’ll go off and dig up the numbers… Stay tuned.

The 8th annual open source convention wraps up today in Portland. I hope to make it there next year. On the open source startup advice front, Matt Asay had an interesting blog post summarizing his presentation. Good stuff Matt and thanks for sharing!

If you go looking for a traditional requisition line category in the Coupa eProcurement Preview Release you will come up empty-handed. It’s not there. Of course, it’s MIA on purpose.

While I was at Oracle I kept trying to figure out how to solve the categorization riddle. On the one hand, classifying spend is a reporting/accounting problem. But on the other, it’s how employees find things. And customers struggle with best practices on how to balance between the two.

So when Oracle introduced it’s eProcurement solution back in the late 90′s, classification was split into 2. You had your findability taxonomy and your reporting/accounting taxonomy. This solved a ton of problems, most notably an upgrade challenge for long-time Purchasing customers. But the maintenance of these two schemes was too much for brand new customers to bear. For R12 (an imminent new release of Oracle Applications), if I’m remembering right, our design was to collapse them back down. But that wasn’t what I really wanted or thought customers needed.

I wanted customers to be able to classify and categorize an item as many times and with as many different “systems” as they felt necessary. But I couldn’t articulate the idea into a comprehensible design. I just kept thinking about changing the data model to say “Reporting Category 1″, “Reporting Category 2″, etc, which was silly.

Now, Noah Eisner deserves the credit for adapting the consumer-centric tags concept to eProcurement – but as we began to dig into how they would work as a replacement for “purchasing category”, I realized I had the functionality I wanted.. And in a Preview release! Who would have thought. The big advance, technically speaking, is dis-associating the category from content & creating an intermediary table to support a many-to-many relationship.

So by all means, think of taggings as classifications of a catalog item (or more generally, content). They can be authored by the Procurement department or by employees in a true decentralized fashion. Or both. The concept is very, very flexible.

Now you’ve probably used tags on YouTube or Flickr. One beef I had was the limit to a single word. I don’t know if people will like our implementation, but it does support use of “double quotes” to encapsulate a phrase.

Tags turn the classification problem over to the user. But leave it in the hands of the Procurement department. And the truth is, it belongs both places.

Now tags don’t just apply to content. So there’s more to cover on this thread another time…

I’m pleased to announce we’ve issued our preview release of Coupa eProcurement. Our corporate website is refreshed, and we are live on Sourceforge. There are still a bunch of things to do to dress up our presence at both places, but it’s a good start.

I want to strongly express my gratitude to Noah Eisner, David Williams, and Seggy Umboh for their impressive dedication, work ethic, and achievement over the last few months. You guys rock. Big time.

Thanks also to Kevin Miller for his early and continuing guidance on the Coupa project. I learned the Procurement space under Kevin’s leadership for many moons at Oracle. He was always a great boss and proved a tough act to follow.

I also want to thank Ron Wohl for his time, his advice, and his continuing encouragement.

Before I go and set the bar too high, please remember Coupa’s preview release is an initial body of work. There are many additional capabilities we plan to add. I’m also looking forward to more detailed feedback from contributors, ISV’s, SI’s, and prospects so we can fine tune and enhance what we’ve done. Still, I think we’ve got some pretty great stuff in there too.

Coupa has now planted its flagpole in the dirt. We’re the Open Source Procurement company.

Wow. Mercury Interactive, a leading provider of testing software (including WinRunner and LoadRunner) is becoming part of HP. You can start to see even more alignment of HP’s businesses towards a “we’re just like IBM” model. What’s next? Selling off the consumer PC business? In any event, HP seems to be doing quite well thank you very much.

I always liked Mercury’s stuff – and I think this acquisition is an interesting one as it will help Mercury recover from their options scandal & immediately stop that stain from being a drag on their business.

In other news Ariba came in light on revenues… Again… But they say things are going really “well” transitioning the business to subscriptions. Of course, that was what they’ve been saying for the last 5 quarters also. I guess it depends on what the definition of “well” is. Sometimes it means on course, good, nicely, etc. But “well” can also mean ‘a deep pit you keep digging in the hope of finding water’… The stock is getting pounded after-hours, down 5.8%. On a more positive note, I really liked Ariba’s recent decision to ship a NetWeaver adapter for their supplier network service to tie into SAP.

As a vendor entering the Procurement space, it sure would be helpful if they starting executing a little better. So hurry up Ariba.

Late last year I began recovering from all the Oracle technology cool-aid I’d read, repeated, and internalized. And as I mentioned previously, I began to recoil from the complexity of Java. Imho some of the IDE’s around Java simply make it worse – a complex wrapper around a verbose language. And before the flame mails roll in from the Java faithful, I’ll pre-empt them by saying a lot of people are doing great things with the language, including the team I admire over at Alfresco. But moving on…

I was surprised how far Microsoft had come with their .Net infrastructure – it’s really good stuff. In fact, hats off to Microsoft on IIS too while I’m at it. (Now why can’t some of the .Net or IIS people work on crappy Internet Explorer – I mean, come on, no transparent png support until IE7?)

Beyond Redmond’s wares, I was surprised to see PHP was now more of a real/serious language too, with much better OO support. Python seemed excellent & gaining momentum also. Many long-time open source advocates predict it will eventually eclipse the other open source scripting languages.

Intuitively (read: I can’t prove it) I began to believe the web applications infrastructure race wasn’t between Microsoft .Net and Java (which admittedly is a horrible apple vs. oranges comparison), but instead between Microsoft .Net and the best open source framework for for developing web applications. And increasingly I believed an object-oriented script language would be a better choice within whichever open source framework won out.

Now, I’m not sure if Ruby is that language. Further, I’m not sure whether Rails will be an open source application framework to rival .Net. But Ruby on Rails has worked incredibly well for us. The productivity gains vs. what I had come to expect at Oracle were shocking and continue to be amazing.

It is clean. It enforces MVC. It’s fairly powerful. We like it a lot.

I first traded emails with David Heinemeier Hannson back in December asking for his advice on whether Ruby on Rails was ready for enterprise software. And I have to say so far it’s been absolutely great. Thank you David!

If you haven’t tried Ruby on Rails or the Ruby language, I’d definitely recommend you give it a shot. And to you Java developers, don’t worry – from what I’ve seen you can make an easy transition to Ruby on Rails. But beware, you may not want to switch back!

After developing Oracle’s eProcurement solution under the leadership of Kevin Miller back in 1997 and 1998 it may seem odd that Noah Eisner and I chose to build our first open source Procurement module in that venerable space. But it’s not. Year in and year out, our experiences led us to believe eProcurement was the place where the “rubber met the road” in terms of effectively managing Procurement across an organization.

And here are some numbers to consider. Worldwide, we estimate eProcurement penetration from Oracle, SAP, Ariba, Ketera, etc, at no more than 7,000 companies. We think that represents live support contracts and subscriptions. Yet many of these support contracts are for implementations that are now inactive – ones that were once started but now abandoned. Many others are “limping along,” never having achieved control over the majority of employee spend.

But even if you assume these 7,000 companies and public sector organizations were delighted with their <insert vendor name here> eProcurement solution, there is still a universe of around 50,000 we think could benefit from an open source solution.

Choosing to deploy an eProcurement solution to your organization is the Football equivalent of “running it up the middle.” It doesn’t get much more basic than promoting self-service buying. It’s useful, important work. And that makes it fun to work on & very rewarding. Noah and I have always enjoyed execution and transaction automation in Procurement over complex optimization and theoretical modeling.

So we thought: wouldn’t it be great to bring the power of effective Procurement to companies of all sizes? And wouldn’t it be great to bring some innovation to a space like eProcurement, which really hasn’t changed that much over the past decade?

As I’ve written previously, I believe Ariba ceded the eProcurement market to SAP and Oracle with their Freemarkets acquisition. It was a good move at the time. They were following the money. Nowadays they still give eProcurement good lip service, but since it is not a revenue driver it can’t get the same focus as their other business areas. Ketera is doing some good work. But the drop-off is dramatic. And so there’s been less innovation in eProcurement than perhaps in an area like Sourcing.

Well, we aim to change that. We hope to offer a true alternative to the Oracle / SAP duopoly rapidly developing in this space. Please join us!

You all may have figured out I am a little wrapped up in something at the moment. But thought you might enjoy this SaaS rant, which I noticed on linkback to Procurement Central. Thanks for sharing your experiences Angus!

Salesforce.com, who Angus comments on, has had a particularly rocky road over the last year on system architecture, splitting instances up to lessen the impact of an outage and distribute load. In many ways, it’s a sign of success. Growing pains. Marc Benioff, Salesforce.com’s CEO and an ex-Oracle executive to boot, seems to be doing an excellent job running that business. Not every company is as socially aware either – check out their Foundation concept.

Angus also mentioned AppExchange, and that business idea is worth keeping tabs on. You can just picture Marc thinking “hey, maybe we can become the business operating system for SaaS…” It’s viral if not open, so who knows… I certainly would never call it “the eBay of business software” but overhyping something so people get the idea isn’t all bad. I hope they choose to break out their AppExchange revenues so we can see if there’s any “there there.” (Getrude Stein quote)

Thought I would post a link to an interesting presentation given by one of the original Lotus 1-2-3 guys, Mitch Kapor. It weighs in at just under 2 hours, so I’d recommend downloading the Google Video client to avoid a painful buffering wait. Here’s Mitch’s bio.

Mitch’s new baby (or newish baby) is called Chandler. It’s a new take on workgroup management software. Find out more here.

In any event, I found the presentation interesting both for Mitch’s commentary on open source as well as his commentary on the history of the software industry. Check it out if you can find the time.

As I noted in an ealier “quickie” post, 5 Procurement SaaS providers sponsored an Aberdeen report finding Procurement SaaS to have advantages over traditional, installed software. And while the report’s findings are obvious based on its sponsors, I liked the report. So by all means, take advantage of the generousity of its sponsors and read it!

The key conclusion was faster time to go-live. Which in turn led to other benefits. And while everyone knows I take issue with some of the hype around SaaS, faster time to go-live is indisputable. It’s a definite advantage.

I was disappointed the only financial analysis presented “for a prototypical $1B company” showcased better ROI for going SaaS simply by the earlier go-live. This smacks of an old IBM study where they said choosing Ariba over an incumbent ERP vendor caused greater ROI because of faster supplier onramping. I think many customers dismiss this type of computation.

There could be other financial projections based on the other claims in the report. Maybe Sudy can produce companion financial case studies at a future date.

The only real negative I could find was in one statistic: 88% of respondents said that in house software system response time & performance was equal or better than SaaS. This reinforces my SaaS = renting-an-apartment analogy. At peak times some customers’ showers are running out of hot water! But it’s a solvable problem – the SaaS providers need to design their on-demand systems for peak load, not average load. And hopefully they will start doing that.

Neither this report nor any of the diatribes of my colleagues has changed my view that SaaS is wonderful but it’s not to be confused with a “category of software.” Instead, it’s a delivery model – one that makes sense where companies want to relieve their already overburdened IT departments and try something out for awhile.

Aberdeen’s data showed the vast majority of firms polled expected to use their SaaS solution 3 years or less (many a whole lot less). In fact, only 16% said they were looking to their “on-demand” provider for a solution for more than 3 years. This is a big red flag for companies that bank their entire business on SaaS – they are betting on a customer base who are just stopping by a while before heading on to their real destinations.

Of course, there’s another way to look at this – enterprises may just pursue a perpetual “shop around” mindset for their SaaS solutions. And SaaS-only vendors might be okay with that – after all, they believe their business model affords them advantages that should become ever clearer over time. I can hear them say it now: “Bring it on!”